Pain During Exercise: Stop, Modify, Monitor or Get It Assessed?
A painful squat, a sore shoulder during a press or an ache after training does not automatically tell you what to do next. Here is how clinical context, exercise demands and the response afterward shape the decision.
Key Facts
✔ A 2017 systematic review included seven randomised trials and 385 people with chronic musculoskeletal pain; painful therapeutic exercise showed a small short-term pain benefit, with no clear medium- or long-term advantage over pain-free approaches [1].
✔ A 2024 patellofemoral-pain best-practice guide synthesised 65 high-quality trials (3,796 participants), patient interviews and clinical reasoning; education and individually selected knee/hip exercise were central [2].
✔ A 2025 multidisciplinary guideline for patellofemoral pain and patellar tendinopathy supports exercise-based care after diagnosis, while grading much of the underlying evidence low or very low certainty [3].
✔ A 2025 review describes delayed-onset muscle soreness after unfamiliar or demanding exercise, often emerging over 24-72 hours; it is not a reliable way to classify all pain after exercise [4].
✔ Victorian public guidance advises people to avoid or modify painful exercise and seek professional advice where needed; guidance for established persistent-pain rehabilitation can differ because the clinical context differs [5, 6].
✔ No single pain number, movement appearance or next-day rule establishes whether a new symptom is safe to exercise through [1, 2, 5].
30-Second Summary
A knee hurts during a squat. Should you finish the set? There is no reliable answer from the word pain alone. Some pain reflects ordinary exercise-related sensations; other pain may require the activity to stop and the person to be assessed.
Research in selected chronic musculoskeletal conditions suggests that exercise need not always be completely pain-free to be useful, but it does not justify pushing through a new injury. A clinician considers when the symptom began, what changes it, how the person moves, what happens afterward and what activity they want back. Stop, modify, monitor and assess are possible decisions, not a universal four-colour safety test.
You are halfway through a set. Now what?
The first few repetitions feel familiar. Then your knee hurts as you lower into a squat. You are not sure whether it is the same sensation you noticed last week, ordinary effort, or something new. The weight has not changed. The decision suddenly feels harder than the exercise: finish the set, reduce the load, walk away or book an assessment?
Most articles offer one of two slogans: listen to your body or no pain, no gain. Both sound decisive. Neither tells you what information you are listening for, or why one person may need to stop while another follows a carefully monitored rehabilitation plan. The useful question is not simply Does it hurt? It is What has changed, and what would that change mean for the next decision?
Pain is real information, but not a complete injury report
Pain is an experience, not a direct readout of tissue damage. The same body part may be painful for different reasons, and pain intensity alone does not identify the structure involved. Equally, that complexity must never be used to dismiss a new symptom. A person cannot infer that an exercise is safe simply because pain and damage do not map perfectly.
Three situations are often grouped under the word pain:
- demanding muscular effort during exercise;
- soreness that appears later after unfamiliar activity; and
- a new or changing symptom that may signal injury or another health issue.
The timing, distribution, function and wider history matter more than forcing each sensation into a simplistic category.
What did the Painful versus Pain-Free exercise research really show?
One of the most useful studies for this question is a systematic review by Smith and colleagues. It examined nine papers from seven randomised trials involving 385 people with chronic musculoskeletal pain. The trials compared rehabilitation exercises that permitted or encouraged some pain with approaches intended to remain pain-free [1].
The researchers found a small short-term improvement in pain favouring the pain-permitting approaches. For medium- and long-term pain, and for function or disability across the assessed time frames, they found no clear advantage of either approach. The evidence base was relatively small, and the paper’s certainty assessments and analyses warrant care: it is not proof that more pain produces better results.
The practical interpretation is narrower, and more useful, than pain is fine. In selected assessed chronic conditions, insisting that every therapeutic movement be completely pain-free may not be necessary for improvement. But neither does a person need to deliberately provoke pain to recover. A new injury, unexplained symptoms, postoperative restrictions or a medical condition may require very different decisions.
This is why two pieces of public advice can sound contradictory. General Australian exercise-safety information tells people to stop or modify painful activity and seek advice [5]. Some persistent-pain rehabilitation resources describe graded exposure and symptom monitoring within an appropriate plan [6]. They are addressing different populations and different levels of clinical knowledge, not offering competing universal laws.
A newer finding: the exercise decision is condition-specific
In 2024, researchers published a best-practice guide for patellofemoral pain, a common presentation involving pain around the kneecap. Their synthesis drew on 65 high-quality randomised trials involving 3,796 people, plus patient interviews and clinical expertise [2]. The conclusion was not everyone should squat this way. It emphasised understanding the person’s symptoms, background, goals and impairments, then using education and appropriately targeted knee or hip exercise. Other approaches were to be selected according to the individual presentation.
A separate multidisciplinary guideline, published online in 2024 and in a 2025 journal issue, also places exercise therapy at the centre of management for diagnosed patellofemoral pain and patellar tendinopathy [3]. Yet it explicitly acknowledges that certainty of some evidence is low or very low. That combination matters: exercise can be a reasonable core treatment while clinicians remain transparent about the limits of the research and the need to tailor decisions.
These are examples from particular knee conditions. They do not establish that all knee pain during squats is patellofemoral pain or tendinopathy, and they cannot be applied as a prescription to an unseen reader.
What about soreness the day after a workout?
A person who tries an unfamiliar hill walk or a new strength exercise may feel relatively comfortable during the activity, then develop muscle tenderness or stiffness later. Delayed-onset muscle soreness (DOMS) is a recognised response to unfamiliar or demanding loading, often developing and peaking in the following one to three days [4].
A 2025 critical review describes DOMS as commonly associated with temporary changes in strength, range and performance. It is not simply “lactic acid stuck in the muscles”, nor does its presence prove that a workout was especially effective [4]. But a delayed symptom is not automatically DOMS. Marked swelling, unexpected loss of function, a new focal injury pattern or symptoms that are severe or progressive require appropriate assessment rather than a reassuring label.
The useful distinction is between timing as a clue and timing as a diagnosis. It started the next morning helps build the history. It cannot, by itself, tell someone that the pain is harmless.
Five questions that change the interpretation
-
When did it start? A sudden new sensation during one repetition raises different questions from familiar symptoms associated with an established condition.
-
What exactly changed? Location, spreading symptoms, unusual weakness, swelling, movement quality and confidence can matter; one pain score cannot capture all of them.
-
What was the total demand? A familiar squat after a long run, poor sleep and an unusually busy week is not necessarily the same demand as that squat on a rested day. This is a clinical reasoning example, not a diagnosis of overload.
-
What happens when the task changes? Load, depth, speed, range, support and volume change the exercise demand. A different response provides information, not proof of the cause.
-
What happens after exercise? Symptoms later that day, sleep, walking, work and next-day function can matter in condition-specific rehabilitation planning. No fixed number of hours is a universal safety clearance.
Four decisions, not a traffic-light diagnosis
-
Stop: If an exercise causes a suspected injury or concerning new symptoms, stopping may be appropriate. Medical emergencies require urgent help. The precise warning signs and escalation wording must be approved by Joy; this article cannot triage an individual.
-
Modify: After relevant assessment, a clinician may alter load, range, tempo, support or volume. This is not permission to keep experimenting with unexplained pain. The point is to change the demand deliberately and observe the response.
-
Monitor: In a pre-agreed rehabilitation plan for a known condition, a clinician may track familiar symptoms during and after activity alongside function. This is different from independently monitoring a new or worsening symptom without assessment.
-
Assess: If pain is new, recurring, changing, unexplained or affecting daily function, the missing information may matter more than another repetition. Physiotherapy assessment can explore movement and physical capacity, with medical referral where indicated.
A squat is not one fixed demand: a practical RheCore demonstration
Consider a person who reports knee pain during a squat. Joy might observe a comfortable chair rise, a supported squat and a version with a different depth or tempo, only where clinically appropriate. The question is not which version looks most aesthetically perfect. It is which variable changes the person’s response and what that might suggest needs further examination.
A higher chair changes the movement range and task demand. Hand support changes balance and how the person distributes effort. Slower movement changes timing and muscular requirements. Additional resistance changes external load. If one version is easier, the finding does not automatically identify weakness, damaged cartilage, a bad knee angle or any specific diagnosis.
This is the difference between giving a generic form correction and making a clinical decision. Observation generates hypotheses. History, examination and response to relevant tests help determine what those hypotheses are worth. A movement that looks symmetrical is not necessarily symptom-free; a movement that looks unusual is not automatically harmful.
The next morning is part of the story, not a pass/fail test
Imagine two illustrative people completing similar sessions. One has an established condition and a familiar response already discussed with their clinician; ordinary function is unchanged afterward. The other develops a new symptom that interferes with walking and worsens the following day. The descriptions are deliberately incomplete. They show why a clinician asks about the after-effect, not which person is safe to train.
Some condition-specific rehabilitation plans use symptom-response rules, but they vary. A 24-hour rule, a two-hour rule or a numerical pain threshold should not be lifted from one condition and turned into advice for everyone. Australian arthritis and persistent-pain resources themselves describe different approaches for different contexts [6, 7].
What does progress look like if pain is not the only score?
A person may be less worried about the exact pain number than about whether they can climb the stairs at home, return to the gym, carry shopping or walk to work. A useful assessment therefore asks about capability as well as symptoms. Can the person perform the relevant task? How much demand can they tolerate? How consistent is the response over time? Has confidence returned?
This does not mean ignoring pain. A worsening symptom still matters even if a strength measure improves. It means measuring both sides: what the person feels and what the person can do. Rehabilitation becomes more meaningful when the progress measure reflects the life the person is trying to get back to.
What information is useful to bring to an appointment?
Instead of arriving with only squats hurt, it can help to describe the activity, when symptoms started, whether they changed during the set, any changes in ordinary function and what happened afterward. The person’s main goal (running, lifting, gardening or simply getting out of a chair) also matters. These observations support a consultation; they do not replace an examination or diagnose the problem.
The RHECORE Perspective
At RHECORE, the point of a movement assessment is not to hand everyone the same rule about pain. It is to make the next decision more informed. A painful squat is an entry point into a set of questions about the person, the movement and its consequences. A change in depth or support may create a useful clue, but the clue is not the diagnosis. Research helps define what is plausible; clinical reasoning helps determine what is relevant to this person.
The destination is not a perfect-looking repetition. It is the capability the person wants back, built with appropriate care, rather than fear or bravado.
Key Takeaways
- Pain during exercise is not a direct measure of injury severity, and it should not be dismissed.
- A seven-trial review found no clear long-term superiority of pain-permitting over pain-free therapeutic exercise in selected chronic musculoskeletal conditions.
- Newer knee-pain guidance supports individually tailored exercise and education after appropriate assessment, not universal squat corrections.
- Delayed muscle soreness and pain arising during a movement are different patterns, but neither timing nor sensation alone makes a diagnosis.
- Load, range, speed, support and total training demands can change the task and provide useful clinical information.
- Stop, modify, monitor and assess are possible decisions, not a public self-clearance algorithm.
- Progress should consider the activity the person wants back, not only pain scores.
References
- Smith BE, et al. Should exercises be painful in the management of chronic musculoskeletal pain? A systematic review and meta-analysis. Br J Sports Med. 2017;51:1679-1687.
- Neal BS, et al. Best practice guide for patellofemoral pain based on synthesis of a systematic review, the patient voice and expert clinical reasoning. Br J Sports Med. 2024;58:1486-1495.
- Dutch multidisciplinary guideline on anterior knee pain: Patellofemoral pain and patellar tendinopathy. Knee Surg Sports Traumatol Arthrosc. 2025;33:457-469.
- Advances in Non-Pharmacological Strategies for DOMS: A Scoping and Critical Review of Recent Evidence. 2025.
- Victorian Department of Health, Better Health Channel. Exercise safety.
- Get Active Victoria. Tips for exercising with chronic pain.
- Arthritis Australia. Physical activity: How will I know if I have done too much?
- healthdirect Australia. How to start exercising.
- Ahpra. Guidelines for advertising a regulated health service.
Frequently Asked Questions
There is no reliable answer from knee pain alone. A new symptom, associated changes in function, the person’s history and the response to activity all affect the decision. If pain is new, unusual or concerning, stop the provoking activity and seek appropriate professional advice.
In selected chronic musculoskeletal rehabilitation trials, some programs permitted pain and achieved outcomes comparable to pain-free exercise over the medium and long term. This does not mean deliberately causing pain is necessary or appropriate for everyone.
Delayed muscle soreness can follow unfamiliar or demanding exercise, but the timing alone cannot establish the cause. New, severe, progressive or function-limiting symptoms need appropriate assessment.
Not necessarily. Load, depth, speed, support, symptoms and personal history all matter. A physiotherapist may compare task variations, but one squat does not diagnose a knee condition.
No single pain number is a universal safety rule. Some diagnosed conditions use clinician-agreed symptom monitoring, but these thresholds should not be applied to a new or unexplained symptom.
Book Now
Was this article helpful?